Where Did the Water Go? A Building-Level Plan for Finding Leaks

A growing water bill can justify an investigation, but it does not identify a leaking pipe. For an Iranian building owner or facility team, the useful investment is a chain from measured volume to a specific inspection and a verified repair. Start with a manageable building or service zone, not a promise to instrument every outlet. The decision is where water is going and who can act when the pattern changes.

EPA’s WaterSense guidance describes submeters as tools for separating tenant or process demand, identifying abnormal use, and tracking improvements. It also describes integration with building management systems. Meter installation is therefore a starting point for operational use, rather than an outcome by itself. [1]

OlbrichCo’s recommendation is to draw the actual boundary on an as-built schematic. Mark incoming utility water, any other supplies, tanks, booster sets, bypasses, zone outlets, irrigation and cooling demand. Identify legitimate unmetered uses and the valve that can safely isolate each noncritical branch. Walk the route with the operator; a drawing alone cannot establish whether a bypass is open or a tenant connection has changed.

Select the first zone by consequence and access: unexplained persistent flow, water-sensitive occupied space, or a major service whose usage can be separated. Agree who pays for investigation, who authorizes an interruption, and who owns the readings after handover. In Iran, specify Persian asset labels and operator instructions, actual replacement availability and a local reading method; assess connectivity and procurement on this project, without assuming a national service condition.

For the proposed boundary, use a volume balance over the same interval: unexplained volume equals all incoming volume minus measured outgoing volume minus estimated legitimate unmetered use minus the increase in stored water. A falling tank level makes that last term negative. This is a conservation-based calculation proposed for diagnosis, not a contractual loss allowance. Include every tank inside the boundary and do not count a transfer between internal tanks as a new supply.

An illustrative calculation, not a project result: during one interval, 12 cubic metres enter, zone outlets record 8, documented unmetered use is 1, and stored volume rises by 3. The residual is zero, not 3 cubic metres of leakage. If storage were unchanged, the same readings would leave 3 unexplained. Even then, investigate clock mismatch, meter uncertainty, estimated uses and missing connections before calling the residual a physical leak.

Record starting and ending tank levels at the meter timestamps and use a verified level-to-volume relationship for that tank’s geometry. Where supply is intermittent, distinguish the refill period from the later drawdown period. With only an upstream meter, a quiet supply line says nothing about water leaving storage. Add a downstream measurement where justified, or use a longer interval with comparable starting and ending storage; document the remaining blind spot.

OIML R 49-1:2024 defines minimum flow Q1 and permanent flow Q3, with specified permissible-error conditions. Its scope is water-volume measurement in a fully charged closed conduit. These distinctions matter when comparing bids for meters intended to reveal small, persistent flows. [2]

PNNL’s metering guidance links selection to the expected demand profile. It warns that oversizing compromises low-flow measurement and notes the trade-off between measuring range and pressure loss. Installation conditions also affect accuracy; the manufacturer’s requirements remain relevant to the chosen instrument. [3]

For procurement, request evidence at the smallest flow the team needs to distinguish, as well as at routine and peak demand. Require the actual model, range, accuracy conditions, pressure loss, temperature rating, orientation and installation clearances. Have the mechanical designer assess full-pipe conditions, air during refill, water characteristics, accessible isolation and any bypass. Do not copy a generic straight-pipe length into every installation or accept a wireless feature as evidence of hydraulic suitability.

Separate operational monitoring from tenant billing. A diagnostic submeter is not automatically acceptable for charging occupants; obtain the applicable approvals and contractual basis before using its figures that way. This note does not assert that OIML adoption, a particular meter type, or submeter billing is mandatory or approved in Iran. The governing requirements, utility conditions and responsible local engineering review must determine those questions.

Commission one complete route from meter register to exported record. Use an agreed safe draw-off or a suitable calibrated reference, within the installed meter’s relevant range, and compare the local increment with the recorded increment. Check litres versus cubic metres, pulse multiplier, direction, timestamps and cumulative-counter reset behaviour. Write the acceptance tolerance before the test using the intended diagnostic resolution and the uncertainty of the measurement chain; do not invent a universal accuracy threshold.

For a site with uncertain connectivity, specify local cumulative totals and enough buffered history for the agreed outage duration. Test a network interruption and recovery without disturbing critical services. Missing data must remain marked missing, never silently become zero demand. Require a readable export containing meter identity, timestamp with time zone, volume, quality flag and event notes. Confirm the operator can retrieve it without the original installer’s account or an unavailable cloud service.

Choose a recording interval short enough to separate the suspected event, but only as short as the meter’s resolution supports. A coarse pulse output cannot establish a precise short-duration flow profile. During commissioning, keep a manual reading and tank-level log alongside the export. Resolve disagreement before creating automated alerts; a clean-looking chart is not evidence that the units, boundary and timing are correct.

DOE’s distribution-audit guidance uses minimum flow during low-use periods as a screening signal and distinguishes a water audit from a survey that locates a leak. It also describes zonal monitoring and specialist leak-location techniques. A screening signal and a repair location are different deliverables. [4]

Establish the zone’s legitimate low-use pattern first: overnight occupancy, cleaning, irrigation, cooling make-up, regeneration cycles and tank refill where present. Compare like operating states. OlbrichCo recommends an alert rule that combines excess flow, duration and an operator-checkable exception list. Set provisional thresholds from observed operation, then test them with a controlled noncritical draw-off. Do not import a single night-flow threshold from another building.

Assign an alert recipient and backup, a response window based on consequence, and a documented route from data check to walkdown, branch investigation and specialist survey if needed. Never let an ordinary water-efficiency alarm automatically isolate fire protection, essential cooling or another critical supply. Any automatic closure needs a separate engineered cause-and-effect assessment, authorized operating procedure and applicable approvals. Escalate visible damage promptly through the site emergency procedure rather than waiting for the next trend report.

EPA’s leak guidance connects occupant reporting with prompt action and recommends repairing leaking equipment according to manufacturer specifications. Monitoring and a reporting channel need a functioning maintenance response to produce an operational benefit. [5]

Close each work order with the location, diagnosed cause, repair, valve restoration and follow-up readings under comparable occupancy, operating and storage conditions. Keep suspected loss separate from confirmed leakage and from estimated avoided consumption. If a savings claim matters commercially, agree the baseline, adjustment method and uncertainty separately; a lower post-repair reading alone is not proof of a quantified annual saving.

Run the first zone through normal occupancy, an actual refill cycle where applicable, and a planned communications test before expanding. Track valid-data coverage, unexplained balance volume with its uncertainty, alerts investigated within the agreed response window, false alarms, confirmed faults repaired, and recurrence after closure. Approve wider rollout only when the team can explain the residuals and complete the response. Include calibration, batteries, spare parts, operator time and service access in lifecycle cost; do not promise an unsupported payback.